From theory to practice - page 407
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
So what? You will have several quotes in a row repeating.
This is the thing based on real ticks:
will be completely distorted.
The peak at zero will go sky-high.
The correct solution is to read the data in a way that maintains maximum consistency with this distribution.
But, not the ticks themselves!!! Two properties must be maintained:
1. correspondence of the number of ticks read to a given timeframe.
2. statistical correspondence to the real tick distribution.
Without histograms it's a bit unclear, I'll post them later for 3 cases of time intervals.
1. uniform (in 2.5 sec)
2. exponential (mean=2.5 sec.)
3. logarithmic (mean=2.5 sec.)
Which of these scales will maximally correspond to the real tick distribution, that one should be used.
This is the most elegant thing you can do.
Each bar will contain approximately the same number of ticks (in the case of a uniform reading, exactly equal). And the distribution of increments will roughly correspond to the real tick.
In this case, you can and should use the "root of T" law for dispersion. The calculations will be as correct as possible.
The most elegant thing will happen.
Each bar will contain approximately the same number of ticks (in the case of a uniform readout, exactly the same). And the distribution of increments will roughly correspond to the real tick.
In this case, you can and should use the "root of T" law for dispersion. The calculations will be maximally correct.
It will be minimally correct, because extremums of BP - the most important characteristic of BP - will be lost.
Will be minimally correct because extremes of BP, the most important characteristic of BP, will be lost.
They may be lost.
But, instead we get the classical analogue (in the case of exponential intervals between ticks within a bar, I really count on it) of the Wiener model.
1. Price experiences chaotic effects within the bar (heavy particle to light particle collisions) - yes
2. Price OPEN or CLOSE of the bar corresponds to uniform measurement time intervals when considering Brownian motion - yes.
Take the formulas for calculating dispersion for diffusion processes and you're done.
Gentlemen!!!
We are closer to the Grail than ever before!!! Don't worry - Uncle Sasha will do everything for you.
Have your pockets ready, please.
"It should be noted that in Forex, tick data does not indicate trades, but price requests. That is, there is one quotation per tick, which does not necessarily result in a trade."
If that were the case, price would often be repeated unchanged in consecutive ticks.
But it isn't - you can look at the ticks yourself, each tick is a price change.
So nonsense is being written by your "guru".
And "price enquiries" were relevant 20 years ago, in negotiated transactions via Reuters Dealing. Now, electronic trading systems, where you can see the price without requests, take up a large share of the market.
does anyone have a seconds quote, for a week at least?
does anyone have a second quote, at least for a week?
CopyTicks: request real ticks for a day or a month.
But, in return we get the classic analogue (in the case of exponential intervals between ticks within a bar, I'm really counting on it) of the Wiener model.
We are closer to the Grail than ever before!!! Don't worry - Uncle Sasha will do everything for you.
Have your pockets ready, please.
Only someone who did not learn mathematics at school can try to make money from the Wiener process.
Since in such a process the next increment does not depend on anything, it is impossible to predict it.
Take care of your pockets, gentlemen. From ignoramuses.
It could also get lost.
But, in return, we get the classic analogue (in the case of exponential intervals between ticks within a bar, I really count on it) of the Wiener model.
In this case, you can and should use the law of "root of T" for dispersion. The calculations will be as correct as possible.
In this case you will get the variance relative to the process value at the starting point of the window, and even then with errors, and not relative to the SMA (which is what we need).